and spring. Population estimates for July, a time when 

 yearlings were dispersing and males were more observable than 

 females, were considered reliable only as a general index to 

 population size. Considering their potential shortcomings, 

 the July estimates were relatively accurate (Table 2.1). 

 Observability of deer during July appeared to increase with 

 population size up to a density of about 3.9 deer/km 2 . 



Observability of deer during autumn surveys also appeared 

 to increase with density to a population of at least 3.9 

 deer /km 2 . Once population level was accounted for, 

 observability was lower during surveys flown with pilot D than 

 with pilot B (Table 2.2). Observability was generally 

 consistent, however, for each pilot. 



The greatest deviations between Lincoln and modeled 

 population estimates for autumn (Table 2.3) occurred in 1983 

 and 1985. Sex and age composition in modeled estimates 

 (Appendix A) had to be consistent with estimates for the 

 previous spring and the subsequent early winter, given 

 natality and mortality during the period. We considered the 

 Lincoln estimate for autumn 1983 inaccurate because 1) bad 

 weather caused a 5-day break in flying mid way through the 

 survey, allowing deer movement between areas, and 2) apparent 

 drought-related movement of some deer from adjacent prairies 

 onto the study area occurred in late summer. The Lincoln 

 estimate for autumn 1985 was considered inaccurate because an 

 unusually early snowstorm resulted in some deer moving from 

 summer to winter ranges during the survey. Overall, the 

 average deviation between the Lincoln and modeled estimates 

 for autumn was 13.1%. When 1983 and 1985 were excluded, the 

 average deviation was 10.0%. Generally, precision for the 

 proportion of marked deer observed during autumn was less than 

 that for winter and spring, even when the same pilot was used. 

 This probably reflected more variable conditions during autumn 

 and the influence of deer density on observability. 



Observability of deer during early winter helicopter 

 surveys was very consistent from year to year (Table 2.4). An 

 average of 66% of all marked deer were observed in surveys 

 flown with pilot A during 1977-1979 and this estimate was 

 precise (CV = 1.8%). During 1980-1986, pilot B was employed, 

 the surveys were flown at an altitude 20-30 m higher than in 

 previous years, and an average 78.8% of the marked deer were 

 observed. Again, the coefficient of variation was low (6.6%). 

 Flying at slightly higher altitude apparently enabled us to 

 observe and record deer which previously had moved before the 

 helicopter over ridges and into areas we had covered earlier. 

 Observability was as high or higher during surveys in years of 

 no snow cover (1981) as in years of complete snow cover 



32 



